v0.6.4 🐞 Bug fix

Fix a bug that cannot set seq_len==n_steps for seq_missing().

What's Changed

• Avoid empty step_idx list when n_steps==seq_len by @WenjieDu in #46
• Release v0.6.4 by @WenjieDu in #47

Full Changelog: v0.6.3...v0.6.4

v0.6.3 🐛 Bug fix

Fixed a bug in MAR that passed arguments are in the wrong order.

What's Changed

• Fix wrong argument order by @WenjieDu in #43
• Release v0.6.3 by @WenjieDu in #44

Full Changelog: v0.6.2...v0.6.3

v0.6.2 refactoring

What's Changed

• Config the project with pyproject.toml by @WenjieDu in #40
• Remove setup files by @WenjieDu in #41

Full Changelog: v0.6.1...v0.6.2

v0.6.1

In this minor update, we enable calc_misssing_rate() to work with pd.DataFrame input.

What's Changed

• Update README, and fix CI testing by @WenjieDu in #36
• Enable calc_misssing_rate work with pd.DataFrame by @WenjieDu in #38

Full Changelog: v0.6...v0.6.1

v0.6 Enable seq and block missing

In this release, we bring sequential and block missing.

What's Changed

• Add seq_missing() and block_missing() by @WenjieDu in #33
• Update seq_missing and block_missing by @WenjieDu in #34
• Releasing v0.6 by @WenjieDu in #35

Full Changelog: v0.5...v0.6

v0.5 Add rdo() to randomly drop observations

In this new version, we add a new missingness-creation function rdo(), which can randomly drop observations only to create missing values.

What's Changed

• Update logo URLs by @WenjieDu in #29
• Update README by @WenjieDu in #30
• Add rdo() by @WenjieDu in #31

Full Changelog: v0.4...v0.5

v0.4

In new version 0.4, we adjust APIs of mcar(), mar_logistic(), mnar_x() and mnar_t(), now they only return the corrupted input data with additional missing values, without masks. Users can use new functions fill_and_get_mask() to get corresponding missing mask and fill the missing data with a specified value like 0.

What's Changed

• Release v0.4 by @WenjieDu in #28

Full Changelog: v0.3...v0.4

v0.3

😎 PyGrinder v0.3

In this release, we fixed some minor bugs and updated docs. Mainly, missingness-creating functions are enabled to only return corrupted X containing NaNs without X_intact and masks by turning the argument return_masks to False.

What's Changed

• Update README by @WenjieDu in #24
• Enable functions to only return corrupted X with NaNs by @WenjieDu in #26
• Fix a minor bug, and release v0.3 by @WenjieDu in #27

Full Changelog: v0.2...v0.3

v0.2

PyGrinder v0.2

This new version is released with below new features:

1. Little's MCAR test;
2. switch the license from GPL-v3 to BSD-3-Clause;

What's Changed

• Adding Little's MCAR test by @WenjieDu in #20
• Switch to the BSD-3 license by @WenjieDu in #21
• Adding pytorch as a dependency lib by @WenjieDu in #22
• Updating the docs and adding pytorch into dependencies by @WenjieDu in #23

Full Changelog: v0.1.1...v0.2

v0.1.1

PyGrinder v0.1.1

What's Changed

• Add two MNAR cases by @AugustJW in #16
• Adding MNAR and MAR functions, and updating the docs by @WenjieDu in #19

New Contributors

• @AugustJW made their first contribution in #16

Full Changelog: v0.1...v0.1.1